SELIMCAN DEDA
SMART BATTERY POWER MANAGEMENT UNIT
Master of Science Thesis
Examiners: Prof. Nikolay T. Tchamov
MSc. Jani Järvenhaara
Examiners and topic approved by the
Faculty Council of the Faculty of
Electrical Engineering on 3 April 2013.
i
ABSTRACT
TAMPERE UNIVERSITY OF TECHNOLOGY
Master’s Degree Programme in Electrical Engineering
DEDA, SELIMCAN: Smart Battery Power Management Unit
Master of Science Thesis, 73 pages, 6 Appendix pages
February 2014
Major: Radio Frequency Electronics
Examiners: Professor Nikolay T. Tchamov and MSc. Jani Järvenhaara
In this project, the ways how a Smart 48V Battery Power Management Unit (PMU)
which uses lead-acid batteries can be designed for wide range of industry usage are
researched.
PMU is a modified version of Battery Management System (BMS). Since the
batteries are designed for charging and discharging for their life, they need an optimum
way to be monitored and/or balanced to increase their efficiency. The PMU system
monitors the voltages and State-of-Charges (SoCs) of whole battery module & individual
cells. PMU ensures reliable and safe battery operations.
MAX11068 is used in the project to meet the monitoring and/or balancing
requirements of the BMS, and thus PMU. However, only using the MAX11068 is not
enough for designing a PMU. In addition to the MAX11068, a microcontroller (Arduino
Due) and a comprehensive Java program are used in the PMU project. The
microcontroller is programmed so that it sends to and receives information from the Java
program. The GUI application is included to let users to investigate the voltage and SoC
variations in whole battery module & individual cells. The users also have an option to
write the data to Microsoft Office Excel sheet. Therefore, they can make use of it
whenever they want.
The complete PMU system monitors the voltages and SoCs of whole battery
module and individual cells accurately. Moreover, the GUI displays the changes the
voltages and SoCs of whole battery module and individual cells on the SoC and voltage
charts.
At the end of the project, the theoretical knowledge is compared with
experimental results.
ii
PREFACE
First and foremost, I would like to thank my supervisor, Professor Nikolay T.Tchamov,
who gave me the opportunity to join the RF Integrated Circuits Laboratory and for his
guidance, advice, encouragement and unwavering enthusiasm throughout my working
time as research assistantship. At many stages of this research project I have utilized from
his experiences to explore new ideas.
I would thank to my examiner Jani Järvenhaara for spending his time to revise my
MSc. Thesis and for his precious comments to make my MSc. Thesis well.
I also would like to thank my dear past and present group members in RF
Integrated Circuits Laboratory for their friendship, help and suggestions to accomplish
my MSc. Thesis project. It is due them that I have enjoyed my working time in RF
Integrated Circuits Laboratory.
Next I would like to address my friends from Finland. I have met so many nice
people from different cultures during my stay in Finland. You made my time so special
and memorable in Finland.
Additional thanks go to my special friends from Turkey and all around world for
providing support and valuable friendship that I need throughout my life.
To Meriç Çağlar, my oldest friend, you are truly more than a special friend, a
brother I call you. Thank you for always giving me motivation to work harder.
I am indebted to my mother for teaching me to hold my head high in life, my
father for giving me advices when I am undecided about anything in life and my family
for their priceless help and support. I am truly gifted to have you as part of my life.
Lastly, I want to dedicate my MSc. Thesis to soul of my dear grandmother who
untimely passed away during the last stage of my MSc. Thesis work and my dear uncle
who passed away long time ago. I believe if they were still alive, they would be proud of
me.
Selimcan Deda
Tampere, February 2014
iii
CONTENTS
List of Figures ..................................................................................................................... v
List of Tables .................................................................................................................... vii
List of Abbreviations ....................................................................................................... viii
List of Symbols ................................................................................................................... x
1.
Introduction ................................................................................................................. 1
2.
Literature Review........................................................................................................ 2
2.1.
What does Battery Management System (BMS) mean? ...................................... 2
2.1.1. The Functions of Power Management Unit (PMU) ...................................... 2
2.1.2. Application Areas of PMU ........................................................................... 3
2.2. Cell Balancing Methods for PMU ........................................................................ 5
2.3.
Lead-Acid Batteries.............................................................................................. 6
2.3.1. The History and Present of Lead-Acid Batteries .......................................... 6
2.3.2. Charging Lead-Acid Batteries ...................................................................... 7
2.3.2.1
Constant Current/Constant Voltage Charging ....................................... 7
2.3.3. Comparison of Rechargeable Batteries ......................................................... 9
2.3.4. Future of Lead-Acid Batteries .................................................................... 11
2.4. State of Charge (SoC) Calculations ................................................................... 11
2.4.1. The SoC Calculation Methods .................................................................... 11
2.5. State of Health (SoH) Estimations ..................................................................... 13
3.
System Considerations Based on MAX11068 .......................................................... 16
3.1.
Inter-Integrated Circuit (I2C) Communication ................................................... 16
3.1.1. Serial Data Line (SDA) and Serial Clock Line (SCL) ................................ 17
3.1.2. I2C Command List ...................................................................................... 19
3.2. MAX11068......................................................................................................... 22
3.2.1. Analog-to-Digital Converter (ADC) ........................................................... 24
3.3. Jumper/Switch Configurations and Cell Connections ....................................... 24
3.4.
Starting up the MAX11068 EV kit .................................................................... 25
3.4.1. MAX11068 EVKit SMBus Ladder ............................................................ 27
3.5. Arduino Due ....................................................................................................... 28
3.6.
Comparison of Selected BMS Boards ................................................................ 30
4.
Power Management System Design ......................................................................... 32
5.
Programming of Power Management Unit ............................................................... 36
5.1.
Flowchart of Programming of PMU .................................................................. 36
iv
6.
5.2.
Java Libraries ..................................................................................................... 38
5.3.
State of Charge (SoC) Calculation Function ...................................................... 38
Results and Discussion ............................................................................................. 39
6.1.
Graphical User Interface .................................................................................... 39
6.2.
Cell Measurement Results.................................................................................. 40
6.2.1. GUI Chart Results ....................................................................................... 42
6.2.2. Excel Results............................................................................................... 44
6.3. Load Discharging Curve .................................................................................... 45
7.
Conclusions ............................................................................................................... 49
8.
Continuation Design of Work ................................................................................... 50
References ......................................................................................................................... 52
Appendix 1: MAX11068 EVKit Cell Configuration Headers and Switch ....................... 57
Appendix 2: MAX11068 EVKit Cell Connection Switches and Headers ....................... 60
Appendix 3: Discharging Enersys Cyclon Batteries ......................................................... 61
Appendix 4: Timing of Parameter Updates During the Relaxation Mode for State of
Health Estimation.............................................................................................................. 62
v
LIST OF FIGURES
Figure 1. General PMU Block Diagram ............................................................................. 3
Figure 2. Some examples of application areas of PMU...................................................... 3
Figure 3. Power (in MW) vs. time (Wikipedia 2009) ......................................................... 4
Figure 4. Cost vs. Increasing Renewable Energy Penetration (Begovic 2012) .................. 5
Figure 5. Charge Current vs. Charging Time for three different current limits (Linden et
al. 2002) .............................................................................................................................. 8
Figure 6. Hyperion AC/DC EOS0606iAD-C battery charger (Hyperion 2013) ................ 9
Figure 7. Enersys Cyclon Battery Open Circuit Voltage vs. State of Charge (Enersys
Cyclon 2008) ..................................................................................................................... 12
Figure 8. I2C-bus bit transmission (NXP Semiconductors 2012) ..................................... 17
Figure 9. START and STOP conditions (NXP Semiconductors 2012) ............................ 17
Figure 10. Complete Data Transfer (NXP Semiconductors 2012) ................................... 18
Figure 11. Functional Communication Initialization Sequence Diagram of MAX11068
(Maxim Integrated 2010) .................................................................................................. 21
Figure 12. Functional Block Diagram of MAX11068 (Maxim Integrated 2010) ............ 22
Figure 13. MAX11068 EVKit (Maxim Integrated) .......................................................... 23
Figure 14. MAX11068 EVKit SMBus Ladder when Board (A) is first board (Maxim
Integrated MAX11068 Evaluation System 2010)............................................................. 27
Figure 15. The front view of Arduino Due (Arduino) ...................................................... 28
Figure 16. The complete measurement system ................................................................. 32
Figure 17. Connection Setup of MAX11068 .................................................................... 32
Figure 18. Series connected Lead-Acid battery design .................................................... 33
Figure 19. Arduino Due & J3 Header of MAX11068 Evaluation Kit+ PACK A
Connection ........................................................................................................................ 33
Figure 20. Load design ..................................................................................................... 34
Figure 21. Complete system setup .................................................................................... 35
Figure 22. The flowchart of programming of PMU ......................................................... 37
Figure 23. PMU GUI without charts................................................................................. 39
Figure 24. Individual cell voltages and SoCs ................................................................... 40
Figure 25. Total Battery SoC vs. Time ............................................................................. 42
Figure 26. Total Battery SoC vs. Time in a different period of time ................................ 42
Figure 27. Total Battery Voltage vs. Time ....................................................................... 43
vi
Figure 28. Total Battery Voltage vs. Time in a different period of time .......................... 43
Figure 29. Instantaneous Battery Voltage, SoC and Status .............................................. 44
Figure 30. Cell voltages and measurement counter displayed in Excel sheet .................. 44
Figure 31. Cell SoC and date & time displayed in Excel sheet ........................................ 45
Figure 32. Total lead-acid battery voltage versus time ..................................................... 46
Figure 33. Load Discharging Curve of lead-acid battery cell #19 .................................... 46
Figure 34. Future BMS/PMU Block Diagram .................................................................. 50
vii
LIST OF TABLES
Table 1. Specifications of Hyperion AC/DC EOS0606iAD-C (Hyperion) ........................ 8
Table 2. Characteristics of different rechargeable batteries (Pistoia 2005) ........................ 9
Table 3. Advantages and disadvantages of lead-acid batteries compared to other
rechargeable battery types (Linden et al. 2002) ................................................................ 10
Table 4. Characteristics of Lead-Acid Batteries (Broussely et al. 2007 & Pistoia 2009). 10
Table 5. Open Circuit Voltage vs. State of Charge of lead-acid battery cell .................... 13
Table 6. Definition of I2C-bus terminology ...................................................................... 16
Table 7. I2C Command List (Maxim Integrated 2010) ..................................................... 19
Table 8. I2C-bus Address Bits (Maxim Integrated 2010) ................................................. 20
Table 9. Jumper Configurations based on number of cells of MAX11068 EVALUATION
KIT+ PACK A & B .......................................................................................................... 24
Table 10. Header P3 Cell Connections of MAX11068 EVALUATION KIT+ PACKA . 25
Table 11. Jumper Configurations of MAX11068 EVALUATION KIT+ PACK A & B . 25
Table 12. Switch Configurations of MAX11068 EVALUATION KIT+ PACK A & B . 26
Table 13. J3 Header Pinout of MAX11068 EVALUATION KIT+ PACK A .................. 26
Table 14. J1 Header Pinout of MAX11068 EVALUATION KIT+ PACK A .................. 27
Table 15. J2 Header Pinout of MAX11068 EVALUATION KIT+ PACK A .................. 28
Table 16. Pin mapping of Arduino Due that used in system ............................................ 29
Table 17. Comparison of Different BMS Evaluation Kits ............................................... 30
Table 18. The used core Java libraries .............................................................................. 38
Table 19. SoC (%) and SoC Icon with respect to it .......................................................... 41
Table 20. Comparison of GUI and multimeter lead-acid battery cell measurements ....... 48
viii
LIST OF ABBREVIATIONS
ACK
ADC
BMS
CC
CC/CV
COM
CV
DC
DoD
EODV
EV
EVKit
FCC
FET
FWUP
GND
GPS
GUI
IC
I2C
I/O
JU
KERS
LDO
LED
LSB
Max
Min
MOS
MOSFET
MSB
NACK
acknowledge
analog-to-digital converter
battery management system
constant current
constant current/constant voltage
communication
constant voltage
direct current
depth of discharge
end of discharge voltage
electric vehicle
evaluation kit
full charge capacity
field-effect transistor
faster wakeup
ground
global positioning system
graphical user interface
integrated circuit
inter-integrated circuit
input/output
jumper
kinetic energy recovery system
low-dropout regulator
light-emitting diode
least significant bit
maximum
minimum
metal-oxide-semiconductor
metal-oxide-semiconductor field-effect transistor
most significant bit
negative-acknowledge
ix
NTC
N/A
OCV
OV
PC
PEC
PMU
PWM
RF
RTD
RX
R/W
SCL
SDA
SMBus
SoC
SoH
SLI
SR
SW
TV
TX
UPS
USB
UV
VDD
VDDBU
VIH
VIL
XOR
negative temperature coefficient
not available
open circuit voltage
overcharge
personal computer
packet-error checking
power management unit
pulse width modulation
radio frequency
resistance temperature detector
receiver
read/write
serial clock line
serial data line
system management bus
state of charge
state of health
starting lighting ignition
repeated start
switch
television
transmitter
uninterruptible power system
universal serial bus
undercharge
positive supply voltage of FET
backup voltage
high-level input voltage
low-level input voltage
exclusive OR
x
LIST OF SYMBOLS
ADC(out)
C(x)
DoD
FCC
ID
PD
Qmax
Qpassed
Qstart
RG
RM
RS
SoC
SoH
Vbat
VBE
output of analog-to-digital conversion
CRC-8 polynomial
depth of discharge
full charge capacity
drain current
power dissipation
maximum battery capacity
passed charge
charge at the beginning
gate resistor
remained capacity
sensing resistor
state of charge
state of health
battery voltage
base-emitter voltage
1
1.
INTRODUCTION
The devices that obtain some or all of their running power from a battery have become
widespread in various industries. The fast enlargement of the use of EVs, portable
electronic devices, remotely operated devices and renewable energy systems generates
high demand for fast propagation of battery technologies (Bergveld et al. 2002).
This project is done because there are only limited applications of smart BMS or
PMU which work with commercially available lead-acid batteries. The main goal of this
MSc. project is to design an accurate and safe PMU monitoring for a certain cell stack
size; including voltage and SoC measurements of battery module and individual cells for
different charging conditions.
In 2011, various German automobile manufacturers proposed using 48V electrical
system standard to obtain higher power than systems using 12V electrical system
standard (Miller 2012). This MSc. Thesis which deals with 48V lead-acid batteries is part
of a larger KERS project (Tampere University of Technology RF Integrated Circuits
Laboratory 2013) that is going to provide energy interaction with DC/DC converter and
motor controller for mobile & stationary big volume applications.
In this MSc. Thesis, the ways how a smart battery PMU can be designed for wide
range of applications are researched. The theoretical background is going to be given
before explaining the PMU research methods and materials. The PMU, a modified
version of BMS, is going to be discussed. The boards which can be used for designing
PMU are going to be compared and the reason of selection of the MAX11068 EVKit is
going to be explained. Moreover, the lead-acid batteries are going to be discussed further.
The importance of SoC and SoH estimations are going to be explained. Thereafter, the
setup process of PMU is going to be stated. The flowchart of smart battery PMU
programming is going to be illustrated. Next, the results of PMU are going to be given
and discussed. After that, the conclusion of PMU project is going to be drawn. Lastly, the
continuation design of work is going to be stated.
2
2.
LITERATURE REVIEW
2.1.
What does Battery Management System (BMS) mean?
According to Andrea (2010), BMS means that; to monitor all cell voltages, protect the
battery cells, estimate the battery’s state of charge, extend battery life, maximize the
battery’s performance, and control the load.
2.1.1.
The Functions of Power Management Unit (PMU)
The PMU, a modified version of a BMS, has crucial power functions in addition to its
battery management functions (Andrea 2010). This is the main difference of BMS and
PMU.
The main function of BMS is to protect all battery cells via keeping them within a
safe operating range. Furthermore, the top balancing (approximately 100% State of
Charge (SoC)) is preferred. In brief, the BMS is an intelligent charging/discharging
system.
Additionally, the battery operated devices and systems have to be managed to
work in their optimum levels. The battery operated devices can be consisted of a large
number of battery cells, and so high voltage levels. For this purpose, the PMU can be
designed to provide optimum use of battery powered devices by maximizing battery
capacity and life, reducing the risks of damaging those devices and controlling the
charging voltage safely. This can be accomplished by controlling (i.e. cell balancing)
and/or monitoring the charging and discharging processes of battery powered devices.
The monitor of PMU can measure and show the figures of merit of conditions of a
battery which are voltage, SoC, State of Health (SoH) and temperature during charging
and discharging based on the requirement of users. However, the monitor of PMU cannot
directly impede the current. In addition to those similar BMS functions, PMU can alter
the effective power-supply source from batteries (Wang et al. 2010). PMU can provide
low power consumption in system without affecting the system functions which separates
it from BMS.
Consequently, PMUs are connectors between batteries and battery operated
devices/systems which enhance battery performance and device/system operation in a
reliable and safe way.
In Figure 1, an example of usage of PMU is given.
2. LITERATURE REVIEW
PC
Microcontroller
3
BMS &
Communication
Interface
Cell Voltage &
SoC &
Temperature
Monitoring
Battery Pack
Current & Voltage
Load
Figure 1. General PMU Block Diagram
In Figure 1, the load can be consisted of motor, DC/DC converters, switches, rectifiers,
power metal-oxide-semiconductor field-effect transistors (MOSFETs) and pumps.
2.1.2.
Application Areas of PMU
PMU has a wide range of applications; from military to medical:
 Automotive Industry (i.e. Electric and Hybrid Vehicles),
 Renewable Energy Systems (i.e. Solar, Wind, Green, and Smart-Sustainable Grid
Renewable Energy Systems),
 Military (i.e. Military Radios),
 Navigators (i.e. global positioning system (GPS)), and
 Portable Medical Systems (i.e. Heart Monitors and Defibrillators).
In Figure 2, four application areas of PMU are shown (forklifts, golf carts,
shavers, and electric vehicles; respectively).
Figure 2. Some examples of application areas of PMU
2. LITERATURE REVIEW
4
Nonetheless, since a smart battery PMU prolongs the battery life, it can be used to reduce
the cost of vehicle during its entire lifetime. Therefore, electric vehicle (EV) batteries get
vehicle manufacturer’s attention. According to Claus, EVs can decrease the gasoline
consumption up to 75% (2008). Moreover, according to the Boston Consulting Group
Inc. (2010), in 2020 the Electric Vehicle batteries market is expected to reach US$25
billion in Western Europe, Japan, United States and China. High quality, reliable, safe,
monitorable and controllable batteries are needed to meet those expectations and to
optimize life of the EVs. Hence, EV manufacturers should consider not only batteries but
also PMUs in rapidly growing EV markets.
Furthermore, the energy storage curves of load power, generated power (with and
without energy storage) and energy flow to storage for the experiment of one day are
given in Figure 3.
Figure 3. Power (in MW) vs. time (Wikipedia 2009)
It is obtained from the Figure 3 that the electrical energy which is produced by
natural resources is varied during the day according to different conditions. It directly
affects the energy storage. However, the energy resources should be adjusted well if there
is no energy storage. For this purpose, PMU can be used for distributed energy storage.
The main challenges of battery energy storage are controlling battery lifetime, power
delivery, providing safety, and handling costs (Rahimi-Eichi et al. 2013). Those
challenges can be completed via PMU.
According to National Geographic (2004), the oil supplies of world are going to
be depleted in 2057. Therefore, battery-powered EVs are going to be needed in future for
2. LITERATURE REVIEW
5
providing required energy. The PMU is going to play a crucial role in controlling and
optimizing energy storage devices and systems in EVs accurately.
Figure 4 shows the flexibility supply curve.
Figure 4. Cost vs. Increasing Renewable Energy Penetration (Begovic 2012)
Figure 4 describes the various approaches of ensuring system flexibility which is
required for defragmenting renewable energy sources. It also shows the variation in
conceptual cost demand of several systems which ensure the integration in electrical
system including supplemental flexibility. The PMU technology may help to manage the
system and decrease risks within this chain.
The improvement of 48V EV motors via PMU with battery fuel gauges provides
some opportunities in electric vehicles industry. The PMU is able to monitor the battery
status and it may control the battery recharging via energy recovery (Pistoia 2005). In
2011, some German automobile manufacturing companies proposed using 48V electrical
system standard to acquire more power than the systems using 12V electrical system
standard while considering 60V electric shock limit of EVs for safety issues (Miller
2012). Therefore, EVs need a leading-edge technology such as PMU.
2.2.
Cell Balancing Methods for PMU
The PMU may protect the battery by stopping the charging current if any cell voltage
exceeds the limit. Moreover, if the temperature also exceeds the limit, then the PMU can
stop the charging current to protect the battery. Similarly, if any cell voltage drops below
the limit, then the PMU may stop the discharging current. For those operations the PMU
uses a well-known function called cell balancing. Cell balancing can be either active or
passive (Andrea 2010).
2. LITERATURE REVIEW
6
When the lead-acid battery cells are connected in series, the current has a direct effect on
all of the cells. If the cell voltages start to differentiate, the charges become unbalanced
which may cause a battery malfunctioning for very long cell series strings at the end.
Hence, a battery balancing mechanism can be used to prevent this potential failure (Krein
et al. 2002).
The capacity of a lead-acid battery can be maximized with cell balancing by
extending two limiting (lowest & highest) points (Arendarik 2012). Those two limits are
defined by one cell in a balanced battery. The charge is removed via preventing the
charging current from the most charged cell and an extra charging current is provided to
weakest charged cell.
In passive balancing the energy is squandered as heat and in active balancing the
energy is transmitted between cells. The passive balancing method of array of cells in
series can be used for lead-acid batteries (Krein et al. 2002). On the contrary, the active
balancing method is not depending on chemical properties of the battery cells. There are
both advantages and disadvantages of passive and active balancing. The passive
balancing method wastes power, but its implementation is easier than the active
balancing method. On the other hand, the active balancing method is faster and the power
efficiency of active balancing method is higher than the passive balancing method, but
the active balancing method is more complex and expensive than the passive balancing
method (Krein et al. 2002).
The cell balancing is optional for series connected 24 battery cells in this smart
battery PMU project. However, hundreds of battery cells totaling hundreds of volts are
required to power electric motors (Miller 2012). If hundreds of series connected battery
cells are used for applications then the cell balancing should be implemented to increase
the efficiency of battery-powered devices and systems.
2.3.
Lead-Acid Batteries
2.3.1.
The History and Present of Lead-Acid Batteries
The first practical lead-acid battery was invented by Gaston Planté in 1860 (Linden et al.
2002) and it was the first rechargeable battery for industrial and commercial use. Planté’s
lead-acid battery cells had a low charging capacity. Following Planté’s inventions, leadacid batteries were developed further and they are being developed now. The lead-acid
batteries are still popular and manufactured globally although they are older than other
rechargeable batteries. Typically, lead-acid batteries are nowadays used in automotive,
electric vehicle, submarine, uninterruptible power system (UPS), television (TV), radio
and alarm systems industries.
2. LITERATURE REVIEW
2.3.2.
7
Charging Lead-Acid Batteries
The recharging process is crucial for batteries to secure the optimum life for them. At
below, three main steps for recharging Lead-Acid batteries are given (Reddy 2011).
1. At the beginning of recharging process, the charging current should not generate
an average battery cell voltage which is greater than 2.4V (gassing threshold). For
cyclic applications, the charging current should not generate an average battery
cell voltage which is greater than 2.5V (Enersys Cyclon 2008).
2. The charging current should be managed so that the cell voltage is kept lower than
2.4V (for cyclic applications, cell voltage is kept lower than 2.5V) during the
recharging process and till 100% of previous discharging capacity is achieved.
3. The current starts to decay to charge finishing rate when the battery begins to
saturate. The recharging process should be ended up at a constant current which is
not more than that charge finishing rate.
There are three main methods for recharging batteries: Constant Current (CC)
charging method, Constant Voltage (CV) charging method or Constant Current/Constant
Voltage (CC/CV) charging method can be used depending on certain conditions.
First of all, since current adjustment is required if the charging current is not
restrained at a low level, the recharging duration becomes longer. Therefore, the CC
charging method is not broadly used for recharging lead-acid batteries. On the other
hand, for charging Enersys Cyclon Lead-Acid batteries the CC/CV charging can be used
because it applies single constant current/voltage level across the battery terminals
regardless of lead-acid batteries’ SoC which are suitable for recharging lead-acid
batteries. Moreover, CC/CV chargers ensure high preliminary current to the lead-acid
batteries due to the high voltage difference between the charger and lead-acid batteries.
According to Enersys Cyclon (2008), constant voltage charging method is
suggested for charging lead-acid batteries when no current limit is required. Therefore,
CC/CV charging method should be applied when a current limit is required.
2.3.2.1
Constant Current/Constant Voltage Charging
According to Linden et al. (2002), modified constant potential with initial constant
current charging method is the fastest and most effective way to recharge lead-acid
batteries. In CC/CV method, a constant voltage with initial constant current is applied to
battery to recharge it. The charging current is at its peak at the beginning of charging
process. Moreover, it is strongly suggested by Enersys Cyclon (2008) that the charging
voltage has to be kept in the range of 2.45V to 2.5V per cell for cyclic applications. If the
2. LITERATURE REVIEW
8
charging voltage is less than 2.45V per cell then there might happen a quick loss in cell
capacity.
In Figure 5, changes in charge current in time are given for 2.5Ah battery charged
by constant voltage of 2.45V. Moreover, chargers are limited to 0.3A, 1A and 2A.
Charge Current(A)
Change in Charging Current vs. time for a 2.5Ah battery charged by constant voltage of 2.45V
2.5
2A
1A
0.3A
2
1.5
1
0.5
0
0
1
2
3
4
5
time(h)
6
7
8
9
10
Figure 5. Charge Current vs. Charging Time for three different current limits (Linden et
al. 2002)
From the Figure 5, it can be observed that the amount of time to recharge the battery is
varied for different current limits. The battery charging process is started with a constant
current for different charging currents. When the battery voltage is reached the gassing
voltage (threshold) the charging current starts to drop. The only difference between those
three charging limits is the amount of time needed to recharge the battery.
In this MSc. Thesis project, Hyperion AC/DC EOS0606iAD-C battery charger is
used. The Hyperion has charge, discharge and storage modes. “Pb” Mode is used to
select lead-acid batteries in Hyperion and charging current is chosen 300mA. 6 seriesconnected cells are charged at the same time due to the specifications. Therefore, total
battery voltage is chosen 12V. The most important specifications of Hyperion AC/DC
EOS0606iAD-C battery charger is given in Table 1.
Table 1. Specifications of Hyperion AC/DC EOS0606iAD-C (Hyperion)
Input voltage range
Range of series-connected lead-acid battery
cells
Charge current
Charge termination
11-15V DC
1-6
100mA ~ 6A max. by 100mA steps for
lead-acid batteries
CV/CC for lead-acid batteries
2. LITERATURE REVIEW
9
The Hyperion AC/DC EOS0606iAD-C battery charger is shown in Figure 6.
Figure 6. Hyperion AC/DC EOS0606iAD-C battery charger (Hyperion 2013)
2.3.3.
Comparison of Rechargeable Batteries
In this part, the characteristics and advantages/disadvantages of lead-acid batteries
compared to other rechargeable battery types are explained.
Table 2. Characteristics of different rechargeable batteries (Pistoia 2005)
Characteristics/Cell Lead-Acid
Specific Energy
30 to 50
(Wh/kg)
Fast Charge Time
8 to 16
(h)
Overcharge
high
Tolerance
Self-Discharge
5
(%/month)
Nominal Cell
2
Voltage (V)
Operating
-20 to 60
o
Temperature ( C)
(only discharge)
Li-ion
150 to 190
Ni-MH
60 to 90
Ni-Cd
40 to 60
Alkaline
80
2 to 3
1
1
2 to 3
very low
low
moderate
moderate
<5
30
20
0.3
3.7
1.25
1.25
1.5
-20 to 60
-20 to 60
-40 to 60
0 to 65
2. LITERATURE REVIEW
10
Hence, it can be observed from the Table 2 that lead-acid batteries have different
characteristics over other rechargeable batteries. According to Linden et al. (2002), the
advantages and disadvantages of lead-acid batteries are given in the Table 3:
Table 3. Advantages and disadvantages of lead-acid batteries compared to other
rechargeable battery types (Linden et al. 2002)
Advantages
Widely used and manufactured globally
Good performance in at low and high
temperatures
High electrical turnaround charge
efficiency (70% ~ 80%) which compares
outward discharge energy with inward
charge energy
High nominal cell voltage (but less than LiIon)
High overcharge tolerance
Explicit SoC demonstration
Less maintenance need
Disadvantages
The energy density (Wh/kg) is low
Hard to be produced in small sizes
Not possible to store in discharged
condition
Wheeled and stationary usage is limited
Thermal runaway can occur if the charging
is not proper
Transportation limitations
Depending on number of cells, the battery
pack might be heavy
According to Broussely et al. (2007) and Pistoia (2009), the characteristics of
lead-acid batteries are given in the Table 4. Note that the Enersys Cyclon D Cell Sealed
lead-acid batteries are used in our PMU project. The recommended EODV (end of
discharge voltage) with respect to the discharge rate is given in table in Appendix 3.
Table 4. Characteristics of Lead-Acid Batteries (Broussely et al. 2007 & Pistoia 2009)
Lead-Acid System
(Application Areas)
Sealed
SLI (starting lighting
ignition) (cars,
aircrafts, ships)
Traction (electric
trucks, road vehicles)
Stationary
(telecommunications,
UPS, energy storage)
Cyclon D Cell
Sealed
Voltage Range
(Float)
(V)
1.8 to 2
1.8 to 2
Operating
o
Temperature( C)
1.8 to 2
-20 to 40
1.8 to 2
-10 to 40
1.93 to 2.14
-65 to 80
-40 to 60
-40 to 55
2. LITERATURE REVIEW
11
It can be concluded from Table 4 that, “Cyclon D Cell Sealed” lead-acid batteries have
different float voltage range than other types of lead-acid batteries. Furthermore, the
operating temperature range of “Cyclon D Cell Sealed” lead-acid batteries is higher than
other types of lead-acid batteries. Cyclon D Cell Sealed lead-acid batteries can be used in
telecommunications, UPS, energy storage, electric vehicles, solar power equipments,
aerospace, etc. Therefore, from all reasons above the “Cyclon D Cell Sealed” lead-acid
batteries are used in our PMU project. In addition, the capacity of Cyclon D Cell SealedLead rechargeable 2V battery is equal to 2.5Ah.
2.3.4.
Future of Lead-Acid Batteries
According to Clarke (2011); cheap expansion, manufacturing scale and 100% recycling
of lead-acid batteries have a key role in fast growing battery markets. In addition, leadacid batteries should satisfy the energy density requirement of Li-ion batteries in future.
Moreover, lead-acid batteries may have less charging time, higher life cycle and more
direct current path to gain advantage over other batteries found in the industry.
Furthermore, the manufacturing process steps and the materials or equipment that
used in lead-acid batteries can be reduced.
2.4.
State of Charge (SoC) Calculations
SoC estimation is a crucial indicator for managing batteries and providing reliable
operations for them. It shows the amount of energy remained in the battery. It prevents
unpredictable and unwanted energy loss of the battery system. It also demonstrates if the
battery cell is overcharged or not. Hence, it provides a longer life for batteries (KongSoon et al. 2008).
2.4.1.
The SoC Calculation Methods
There are different ways to calculate SoC of lead-acid batteries. One solution method for
lead-acid battery SoC estimation is presented by Pang et al. (2001). The approach of Pang
et al. (2001) to calculate SoC of lead-acid batteries is based on estimating the open circuit
voltage (OCV) of the lead-acid battery and then providing a linear relationship between
OCV and SoC.
According to Pang et al. (2001):
OCV  
SoC 
(1)

where β is the lead-acid battery terminal voltage when SoC is equal to 0% (when leadacid battery charge is drained) and α is obtained via putting the OCV and β values in
equation (1) when SoC is equal to 100% (when lead-acid battery is fully charged).
2. LITERATURE REVIEW
12
Additionally, the relation between Cyclon cell voltages and SoC is based on the Figure 7.
According to Enersys Cyclon (2008), SoC of Enersys Cyclon battery can be calculated
using the curve given in the Figure 7.
Cyclon Battery Open Circuit Voltage vs State of Charge
2.15
Open Circuit Voltage (V)
2.1
2.05
2
1.95
1.9
10
20
30
40
50
60
70
State of Charge (%)
80
90
100
Figure 7. Enersys Cyclon Battery Open Circuit Voltage vs. State of Charge (Enersys
Cyclon 2008)
The Figure 7 is valid to within 20% of correct SoC of battery cell, if the cell has not been
applied any charge/discharge operations within the last 24 hours. In addition, the Figure 7
is valid to within 5% of correct SoC of battery cell, if the cell has not been applied any
charge/discharge operations within the last 5 days. Therefore, the open circuit voltage
(OCV) of lead-acid batteries is dependent of charging/discharging rates and the amount
of time they have been disconnected from charger.
The Cyclon battery should have the required overcharge to reach long battery life.
Under proper charging conditions the Figure 7 is valid for ten years (Enersys Cyclon
2008). The proper charging method for Cyclon batteries is Constant Current/Constant
Voltage (CC/CV) method which is explained in section 2.3.2.1.
Furthermore, it is observed from the Figure 7 that the SoC of lead-acid battery is a
linear function of open circuit voltage. The battery cell is considered as fully charged
when the OCV is greater than 2.14 V/cell and fully discharged when the OCV is less than
1.94 V/cell. The relationship between OCV and SoC of a lead-acid battery cell is given in
the Table 5.
2. LITERATURE REVIEW
13
Table 5. Open Circuit Voltage vs. State of Charge of lead-acid battery cell
OCV (V)
>2.14
(2.13-2.14]
(2.12-2.13]
(2.11-2.12]
(2.1-2.11]
(2.09-2.1]
(2.08-2.09]
(2.07-2.08]
(2.05-2.07]
(2.04-2.05]
(2.03-2.04]
(2.02-2.03]
(2.01-2.02]
(2-2.01]
(1.99-2]
(1.98-1.99]
(1.97-1.98]
(1.96-1.97]
(1.95-1.96]
(1.94-1.95]
<1.94
SoC (%)
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
According to the Table 5, the lead-acid battery cell reaches 100% SoC if the
voltage level of that cell is greater than 2.14V.
2.5.
State of Health (SoH) Estimations
Battery SoH is an important estimation to observe the battery performance. According to
Tanaami et al. (2009), SoH estimations show how much the battery can supply its
maximum capacity under usual charging or discharging conditions. The battery aging
comes up when the charge and discharge cycles of a rechargeable battery are increased.
In addition, the battery aging can occur if the battery is used improperly during charging
or discharging. The battery aging is an important reason to affect battery life and
capacity.
According to Huet (1998), the charging or discharging current magnitude, DoD
(depth of discharge), room temperature, battery charge control method, over charging or
discharging, storage type and duration are the factors that affect the battery SoH.
An early breakdown of the lead-acid battery of the device can cause a problem
since the electrical power consumption of that device is increasing by the time of
progress. Hence, the battery SoH estimation is crucial and required for devices during the
2. LITERATURE REVIEW
14
process. The Gas Gauge Algorithm can be used to estimate SoH. The equations of Gas
Gauge Algorithm to estimate SoH are given in following equations step by step (Texas
Instruments 2006).
(2)
DoD  1  SoC
Simply, the depth of discharge given in the equation (2) is the complement of
state of charge. The gas gauge updates data on DoD based upon OCV (open circuit
voltage) readings during a relaxed mode for each cell individually (Texas Instruments
2006). As it is illustrated in figure given in Appendix 4, the subsequent DoD
measurements are taken when 30 minutes elapsed in the relaxation mode. After that DoD
measurements are taken every 100 seconds.
Qpassed  Q1  Q2
(3)
where Q1 is the first passed charge from the fully charged state and similarly Q2 is
the second passed charge from the fully charged state. Qpassed is set to zero for every
DoD0 update.
Qpassed
Q max 
(4)
SoC1  SoC 2
where
SoC1 
Q1
Q max
(5)
and similarly
SoC 2 
Q2
Q max
(6)
In the equation (4), Qmax denotes the maximum capacity of the battery has.
Qmax is updated when:
dV
(7)
 4V / s
dt
The equation (7) denotes the rate of instantaneous voltage change over time (volts
per second). As it is obtained from the figure given in Appendix 4, Qmax is also updated
when the maximal time is exceeded.
(8)
RM  ( DoDfinal  DoDstart ) xQ max
Thus, the Gas Gauge Algorithm utilizes DoD and Qmax (total chemical capacity)
informations to estimate the RM (remaining capacity).
(9)
FCC  Qstart  Qpassed  RM
where Qstart denotes the passed charge which makes DoD equal to DoD0. Qstart is
equal to zero for a fully charged battery. DoD0 stands for the last DoD reading before
charging or discharging. FCC (full charge capacity) is the passed charge quantity from
fully-charged condition to the system termination voltage. The passed charge must be
higher than 37% of design capacity to allow the occurrence of an update. In other words,
2. LITERATURE REVIEW
15
FCC is the available capacity at the present charging or discharging rate. FCC depends on
the temperature at a time of measurement and load.
Consequently, the formula to calculate SoH is given in the equation (10)
(Williams 2012):
FCC
(11)
SoH 
Design Capacity
In brief, the SoH monitors the chemical degradation and shows when the battery
should be replaced.
15
3.
SYSTEM CONSIDERATIONS BASED ON
MAX11068
The system considerations given in this chapter describe MAX11068 which is used in
this Smart Battery Power Management Unit (PMU) project.
3.1.
Inter-Integrated Circuit (I2C) Communication
I2C is a multi-master bi-directional 2-wire bus for efficient inter-IC control. Multi-master
means more than one device capable of controlling bus can be connected to it. However,
only one master is active at any time. It maximizes hardware efficiency and circuit
simplicity. Some of crucial features of I2C are listed below (NXP Semiconductors 2012):






Only serial data line (SDA) and serial clock line (SCL) are required and it is
exactly software-defined.
All I2C-bus compatible devices communicate directly with each other via I2C-bus.
Masters can run as master-transmitters or as master-receivers. They control the
I2C-bus by sending data and addresses.
It has collision detection, clock synchronization and arbitration properties to block
data overlap if two or more masters simultaneously begin data transfer.
Permits a system to be changed easily by clipping or unclipping ICs (integrated
circuits) to or from the bus without affecting any other functions of I2C-bus.
The I2C-bus interfaces are not required to be changed.
SDA and SCL convey the information between devices that connected to I2C-bus.
The devices that connected to I2C-bus have their unique addresses. More importantly, for
the PMU projects I2C communication provides very low current consumption, high noise
independence and broad range of the supply voltage and temperature.
Table 6. Definition of I2C-bus terminology
Term
Transmitter
Master
Synchronization
Arbitration
Task
Sends data to I2C-bus
Begins a transfer, creates clock signals and ends a transfer
Synchronizes clock signals of more than one device
If more than one master simultaneously attempts to control I2C-bus,
only one is supposed to do so the other master loses the arbitration
and waits until it is allowed to begin transmission
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.1.1.
17
Serial Data Line (SDA) and Serial Clock Line (SCL)
According to the NXP Semiconductors (2012), the SDA and SCL are both bidirectional.
When the I2C-bus is not assigned any job, the SDA and SCL levels are both HIGH (‘1’).
Since any device which may have metal-oxide-semiconductor (MOS) and bipolar
technologies can be connected to I2C-bus, the SDA and SCL levels are not defined as
LOW (‘0’) or HIGH (‘1’). The levels depend on positive supply voltage of FET (VDD).
The low-level input voltage (VIL) is set as 0.3VDD and high-level input voltage (VIH) is
set as 0.7VDD.
SDA
SCL
Data line
is stable
Data
change is
permitted
Figure 8. I2C-bus bit transmission (NXP Semiconductors 2012)
From Figure 8 it is obtained that the data on SDA has to be stable when the clock
signal is HIGH. Moreover, the LOW or HIGH level of SDA is changed when the SCL
clock is LOW. In addition, for each transmitted data bit one clock pulse is produced.
SDA
SCL
START
Figure 9. START and STOP conditions (NXP Semiconductors 2012)
STOP
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
18
All transmissions that are generated by the master begin with START and end with
STOP. As observed from Figure 9, for SDA; the START is acquired by a HIGH to LOW
passing when SCL is HIGH. On the other hand, for SDA; the STOP is acquired by a
LOW to HIGH passing when SCL is HIGH.
Each byte is followed by an acknowledge (ACK) bit. Data transfer is begun by
most significant bit (MSB) first. The ACK bit demonstrates that the data byte
transmission is done successfully. When ACK is sent, the receiver sets the SDA as LOW.
However, the ACK bit does not ensure the data integrity. The negative-acknowledge
(NACK) bit is generated if there is no receiver on I2C-bus or the receiver is not able to
receive any data byte. When NACK is sent, the receiver sets the SDA as HIGH.
SDA
SCL
START
1-7
8
9
ADDRESS
R/W
ACK
1-7
8
DATA
9
ACK
STOP
Figure 10. Complete Data Transfer (NXP Semiconductors 2012)
In Figure 10, the slave address is sent after START. The slave address is sevenbits long. It is followed by the read/write (R/W). R/W is a data direction bit. If R/W is
zero then the master writes to slave. On the other hand, if R/W is one then the master
reads from slave. The data byte transmission is ended by STOP which is generated by the
master.
For example, when the master generates a Write, it sends data bits and receives an
ACK bit.
If the slaves are slower than masters then the slaves can spread the clock pulse via
setting SCL low in a short span of time until it is able to make a new data transmission.
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.1.2.
19
I2C Command List
In the Table 7, the commands together with their functions are given. If the device
configuration is addressable, then the initializations given in Table 7 is not used. In
addition, Table 7 is specific for the MAX11068.
Table 7. I2C Command List (Maxim Integrated 2010)
Name
HELLOALL
ROLLCALL
SETLASTADDRESS
WRITEALL
WRITEDEVICE
READALL
Function
Setting the all device
address via initializing the
first device’s address in the
chain. All other remaining
devices are autoincremented.
Determining # of devices in
the chain.
Determining which device
address is the last one.
Enabling a value to be
written to a register in all
active devices at the same
time.
Writing the data to only a
certain device.
Reading the data from a
certain device register for
all devices in the chain.
PEC Byte
Not contain
Not contain
Required from the host
Required from the host
Required from the host
Send to the host
According to Maxim Integrated (2010), the first two bits of HELLOALL have to be ‘1’.
Next 5 bits determine the address of first device in system management bus (SMBus)
ladder. The last bit is the R/W bit and it must be ‘0’. The 5-bits starting address is
determined as the least significant bit (LSB) is the first bit. The typical starting address is
0x01. Therefore, the HELLOALL command becomes 11100000 in binary format. It is
equal to 0xE0 in hexadecimal format.
The ROLLCALL command should be called just after HELLOALL command.
When 0xFF is returned it means that the host has addresses of all devices. The
ROLLCALL command is a read of address 0x01. The write bit is added at the end of 7bits broadcast address and it is sent on I2C-bus. Next to this data, the command of address
0x01 register is sent.
The SETLASTADDRESS command is called after ROLLCALL command. The
write bit is added at the end of 7-bits broadcast address and it is sent on I2C-bus. Next to
this data, the command of address 0x01 register is sent. After that two data bytes are
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
20
written to each device which is sent on I2C-bus. The first data byte may be any value.
However, the second data byte is written to address register of all devices. Thereafter the
PEC (packet-error checking) byte is sent. PEC byte is calculated from first four bytes.
The WRITEALL command is called after SETLASTADDRESS command. The
write bit is added at the end of 7-bits broadcast address and it is sent on I2C-bus. Next, the
command of address is sent with MSB first. The lower and upper data bytes are added to
the data, respectively. After the PEC byte is sent. The PEC byte is calculated from first
four bytes.
The WRITEDEVICE command is called after WRITEALL command. The only
difference of WRITEDEVICE and WRITEALL command is the address byte of I2C-bus
of WRITEDEVICE command has fixed MSBs (0b10 in binary format), pursued by the
command of address is sent with LSB first in the place of broadcast address.
The READALL command is called after WRITEDEVICE command. The write
bit is added at the end of 7-bits broadcast address and it is sent on I2C-bus. Next to this
data, the command byte is sent. To change the direction of bit stream, the repeated start
(SR) is sent. On the contrary, this time the read bit is added at the end of 7-bits broadcast
address instead of the read bit and it is sent on I2C-bus.
In Table 8, the address bits of commands are shown. The Broadcast Address is
used for ROLLCALL, WRITEALL and READALL commands.
Table 8. I2C-bus Address Bits (Maxim Integrated 2010)
Command
Broadcast Address
(Default)
HELLOALL
WRITEDEVICE
7
B7
0
1
1
6
B6
1
1
0
5
B5
0
A0
A0
4
B4
0
A1
A1
3
B3
0
A2
A2
2
B2
0
A3
A3
1
B
0
A4
A4
R/W
1/0
1/0
0
0
In the Smart Battery PMU project the PEC bytes are calculated via Pseudo-Code
Algorithm for a CRC-8 PEC Calculation (Maxim Integrated 2010). MAX11068 utilizes
SMBus PEC algorithm to sustain the completeness of data. The CRC-8 algorithm
supports SMBus PEC mechanism to carry out calculations. The CRC-8 polynomial
(C(x)) is shown below:
C ( x)  x8  x 2  x  1
At first, C(x) is initialized to zero. The byte is XORed (exclusive ORed) with C(x)
for each input byte and called as ‘Remainder’. When MSB is equal to one then the
remainder is left-shifted and XORed with eight LSBs of the C(x). When MSB is equal to
zero then the remainder is left-shifted by one bit. This is applied until eight left-shifts are
acquired. Next, the transaction is repeated on the next input byte using current C(x).
When the process of all input bytes is completed the last result is the output byte of C(x).
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
Consequently, MAX11068 gets functional after any reset event. In Figure 11, the
complete functional I2C SMBus ladder communication initialization sequence diagram of
MAX11068 is shown. The SMBus ladder modules are set for operation after they are
regulated for communication.
MAX11068
HELLOALL
NACK
COMMUNICATION ERROR
ACK
ROLLCALL
NACK
COMMUNICATION ERROR
ACK
SLAVE DATA IS RECEIVED
UNTIL CONSECUTIVE 0xFF
BYTES, THEN SLAVES ARE
NACKED.
CALCULATION OF
NUMBER OF DEVICES
IS CALCULATION OF
NUMBER OF DEVICES
CORRECT?
NO
SYSTEM ERROR
YES
SETLASTADDRESS
NACK
COMMUNICATION ERROR
ACK
STATUS REGISTER
BITS ARE CLEARED
COMMUNICATION IS
INITIALIZED
Figure 11. Functional Communication Initialization Sequence Diagram of MAX11068
(Maxim Integrated 2010)
21
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.2.
MAX11068
The MAX11068 integrated battery sensor is used for managing the lead-acid batteries in
this MSc. Thesis. The BMS topology is master-slave architecture in MAX11068. The
MAX11068 can measure the voltage of up to 12 cells and has two auxiliary ports which
can be used with thermistors for temperature measurements. Moreover, the MAX11068
evaluation kit (EVKit) can measure the voltage of up to hundreds of cells depending on
the battery selection. In Figure 12, Functional Block Diagram of MAX11068 is illustrated
(Maxim Integrated 2010).
DCIN
VAA
REF
GNDU VDDU
Precision
+2.5V
Reference
I2C Upper
Port
6V to 72V
LDO Linear
Regulator
THRM
3.4V
ALRMU
SCLU
SDAU
POR
AUXIN2
AUXIN1
DIS Select
(11:0)
6 MHz
Oscillator
C12
C11
C10
+
C9
Switch Bank
C8
-
C7
C6
C5
C4
Level Shift
Instrumentation
Amp
12-bit ADC
CP+
CP-
Control and
Status
Cell Equalization
COM
3.4V
C3
C2
C1
32 kHz
Oscillator
ALRML
C0
I2C Lower
Port
3.4V
SCLL
SDAL
SHDN
SW Select
(26:0)
AGND
GNDL VDDL
GPIO0 GPIO1 GPIO2
Figure 12. Functional Block Diagram of MAX11068 (Maxim Integrated 2010)
In Figure 12, cell data is obtained by consecutive approximation of 12-bit analogto-digital converter (ADC). It also enables to measure internal and external temperature.
The switch bank results from a high voltage multiplexer. It forms switching between the
cells. When cells are switched; the cells are measured by the 12-bit ADC and after that
they are stored in internal memory. All cells are measured with 2.5V voltage reference
for gaining high accuracy. Internal switch can be used to charge a cell slower than other
cells if that cell has higher voltage than the other cells. The whole MAX11068 is
controlled by the 6MHz internal oscillator. It determines the communication speed and
ADC operations. I2C Upper Port (master port) is level shifted and referenced to ground
(GNDU).
22
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
MAX11068 is a programmable battery-monitoring data collection IC whereas
MAX11080 is a battery pack fault monitor IC. The MAX11068 monitor measures battery
parameters (voltage, temperature, State of Charge (SoC), State of Health (SoH), etc.)
including the control loop. It controls indirectly the charger and the load. A monitor
cannot interrupt the pack current (Maxim Integrated 2010). The balancer maximizes the
pack performance via providing cell balancing. The balancer is sufficient for the BMS as
long as it can control the charger and the load. The MAX11080 IC only checks for
under/over voltage and temperature. It cannot monitor the SoC or SoH. The fault signal
of the MAX11080 is not used in this MSc. Thesis. From Figure 13, the MAX11068
EVKit can be seen.
Figure 13. MAX11068 EVKit (Maxim Integrated)
The numbers given in Figure 13 denotes that:
1.
2.
3.
4.
5.
6.
7.
Arduino Connection Header
Auxiliary Inputs
Header that is connected to battery cells
Cell configuration switch
Ribbon cable that connect two MAX11068 EVKits
Switch that switches the battery cells in and out of BMS
Cell stack voltages
23
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.2.1.
24
Analog-to-Digital Converter (ADC)
The MAX11068 can measure 12 cells. A 12-bit ADC is utilized to digitize the cell
voltages. The MAX11068 measures each cell differentially. The MAX11068 has a
measurement range from 0 to 5V (Maxim Integrated 2010).
The output of ADC shows the cell measurement result. The following equation is
used to calculate the cell voltage.
ADC (out )  5
CellVoltage 
4096
In the equation above, the 4096 comes from 212 since the maximum data bits are
equal to 12. In addition the ADC(out) is multiplied by 5 because the maximum cell
voltage measurement could be equal to 5V.
3.3.
Jumper/Switch Configurations and Cell Connections
Table 9 provides the complete shunt positions for the MAX11068 evaluation kit (EVKit).
The Tables 9 and 10 are done according to Maxim Integrated MAX11068 Evaluation
System (2010).
Table 9. Jumper Configurations based on number of cells of MAX11068 EVALUATION
KIT+ PACK A & B
# of
Cells
2
3
4
5
6
7
8
9
10
11
12
Position of Shunt
JU2,
JU102
JU3,
JU103
JU4,
JU104
JU5,
JU105
JU6,
JU106
JU7,
JU107
JU8,
JU108
JU9,
JU109
JU10,
JU110
JU11,
JU111
On
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
On
On
Off
Off
Off
Off
Off
Off
Off
Off
Off
On
On
On
Off
Off
Off
Off
Off
Off
Off
Off
On
On
On
On
Off
Off
Off
Off
Off
Off
Off
On
On
On
On
On
Off
Off
Off
Off
Off
Off
On
On
On
On
On
On
Off
Off
Off
Off
Off
On
On
On
On
On
On
On
Off
Off
Off
Off
On
On
On
On
On
On
On
On
Off
Off
Off
On
On
On
On
On
On
On
On
On
Off
Off
On
On
On
On
On
On
On
On
On
On
Off
As 12 cells are used for each MAX11068 Evaluation Kit+ Pack in our Smart
Battery PMU project, the shunts given in Table 9 should be positioned off. Moreover, the
cell stack voltage is ensured by cascading 12 battery cells between C0_A to C12_A and
C0_B to C12_B nodes. In addition to Table 10, for MAX11068 EVALUATION KIT+
PACKB; header P103 has same cell connections as P3.
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
25
Table 10. Header P3 Cell Connections of MAX11068 EVALUATION KIT+ PACKA
CELL #
1
2
3
4
5
6
7
8
9
10
11
12
3.4.
+TERMINAL
P3-2
P3-4
P3-6
P3-8
P3-10
P3-12
P3-14
P3-16
P3-18
P3-20
P3-22
P3-24
-TERMINAL
P3-4
P3-6
P3-8
P3-10
P3-12
P3-14
P3-16
P3-18
P3-20
P3-22
P3-24
P3-26
Starting up the MAX11068 EV kit
After applying at least 6 Volts across MAX11068 EVALUATION KIT+ PACKA+ and
PACKA-, SHDNINA net should be droved high. In our project, we control SHDNINA
by our software. To do this, in addition to the Table 10, the jumpers and switches should
be configured as in the Table 11 according to the Maxim Integrated MAX11068
Evaluation System (2010).
Table 11. Jumper Configurations of MAX11068 EVALUATION KIT+ PACK A & B
Jumper
JU0, JU100
JU1, JU101
JU12, JU112
JU13, JU113
JU14, JU114
JU15, JU115
JU16, JU116
JU17, JU117
JU18, JU118
JU19, JU119
JU20, JU120
JU21, JU121
JU24, JU124
JU25, JU125
JU26, JU126
JU28
JU128
JU31, JU131, JU32, JU132
Position of Shunt
On
On
Off
1-2
2-3
2-3
1-2
2-3
1-2
1-2
On
Off
On
On
On
2-3
1-2
On
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
26
Jumpers JU0, JU1, JU100 and JU101 enable or disable the auxiliary inputs to measure
external resistance temperature detector (RTD) components (Maxim Integrated
MAX11068 Evaluation System 2010). The negative temperature coefficient (NTC) RTD
can be structured with auxiliary analog inputs to measure battery cell temperature. Since
the temperature measurement of battery cells is not a part of our PMU project, the
auxiliary inputs are not used via installing a shunt on jumpers JU0, JU1, JU100 and
JU101. The cell configuration switch SW101 is given in Appendix 1 and the cell
connection switches are given in Appendix 2.
Table 12. Switch Configurations of MAX11068 EVALUATION KIT+ PACK A & B
Switch
SW1, SW101
SW21, SW22, SW121, SW122
Switch Position
Off
On
After the all configurations, using a universal serial bus (USB) cable; the MAX11068
Evaluation Kit+ Pack A is connected to personal computer (PC) via Arduino Due.
Arduino Due is connected to MAX11068 Evaluation Kit+ Pack A’s J3 header.
Table 13. J3 Header Pinout of MAX11068 EVALUATION KIT+ PACK A
Pin Number
1
Net Association
3
5
7
9
11
SDALA
GNDLA
SCLLA
Not Configured
13
15
17
2, 4, 6, 8, 10, 12, 14, 16, 18, 19
Not Configured
Not Configured
Not Configured
GNDLA
SHDNINA
SHDNA
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.4.1.
27
MAX11068 EVKit SMBus Ladder
MAX11068 Evaluation Kit utilizes I2C communication to cascade up to 31 MAX11068
devices via routing I2C communication to headers J1 & J2 for Board (A) and J101 &
J102 for Board (B). Since one shipped MAX11068 Evaluation Kit involves two boards
(boards A & B), 16 MAX11068 Evaluation Kits are required to acquire 31-device
SMBus ladder (Maxim Integrated MAX11068 Evaluation System 2010).
In Figure 14, the MAX11068 EVKit SMBus Ladder can be found when Board (A) is the
first board (Maxim Integrated MAX11068 Evaluation System 2010).
J1
J2
(A)
J3
J101 J102
J1
(B)
J103
J2
J101 J102
(A)
(B)
J3
J103
Figure 14. MAX11068 EVKit SMBus Ladder when Board (A) is first board (Maxim
Integrated MAX11068 Evaluation System 2010)
In the Smart Battery PMU project two MAX11068 boards (boards A & B) are used. They
are connected via ribbon cable that placed across the jumpers J2 and J101. Furthermore,
the cell configuration headers J101 and J102 are given in Appendix 1.
Table 14. J1 Header Pinout of MAX11068 EVALUATION KIT+ PACK A
Pin Number
1
Net Association
3
5
7
9, 10, 11, 12
13
SCLLA
SDALA
ALRMLA
C0_A
SHDNINCA
15, 17
19
2, 4, 6, 8
14, 16, 18, 20
Not Configured
ALRMOUTCA
GNDLA
AGNDCA
SHDNA
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
Table 15. J2 Header Pinout of MAX11068 EVALUATION KIT+ PACK A
Pin Number
1
3
5
7
9, 10, 11, 12
13
15, 17
19
2, 4, 6, 8
14, 16, 18, 20
3.5.
Net Association
CP+A
SCLU_A
SDAU_A
ALRMU_A
C12_A
CP+CA
Not Configured
ALRMINCA
GNDU1
GNDU2
Arduino Due
The Arduino Due is chosen for this MSc. Thesis. Arduino Due is a microcontroller board
which has Atmel SAM3X8E 32-bit ARM chip (Atmel) on it. The Arduino Due has 54
digital pins and each of them can be utilized as an input or output. Nonetheless, 12 of
digital I/O pins can be used as pulse width modulation (PWM) outputs. They operate at
3.3V. The input/output (I/O) pins can tolerate maximum 3.3V. Additionally, the direct
current (DC) current for 3.3V pin is 800mA. The programming port of Arduino Due is
connected to an ATmega16U2 microcontroller which ensures a virtual communication
(COM) port (Arduino).
The Figure 15 shows the front view of Arduino Due. The Arduino Due is connected to
PC by micro-USB cable. When data is being transmitted between the PC and Arduino
Due, the receiver (RX) and transmitter (TX) light-emitting diodes (LEDs) start to flash
(Arduino).
Figure 15. The front view of Arduino Due (Arduino)
28
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
The MAX11068 already includes the USB-powered command module MINIQUSB
together with the EVKit software. The MINIQUSB provides the communication between
the MAX11068 EVKit and PC (Maxim Integrated 2008). However, the MINIQUSB and
EVKit software system is not modifiable and an improved version of visualization is
required for user-friendliness. Therefore, a new modifiable system is needed. The
Arduino Due is used in order to satisfy that need. The Arduino Due is programmed based
on the Maxim Integrated (2010). The MAX11068 EVKit becomes controllable with the
Arduino Due. In addition, a new visual quality is provided via a Java graphical user
interface (GUI) which is going to be discussed in “Results and Discussion” chapter.
The Arduino Due mainly controls, measures and records the data. Therefore, the
title of whole Battery PMU project is extended as ‘Smart’ Battery PMU. The
microcontroller board interfaces with the battery packs via I2C communications.
When the battery cells are charged off-limits, the microcontroller has a feature to
switch off the charging process. On the other hand, it can charge the weakest battery cell
to provide cell balancing. It can be designed to provide a long-life PMU system.
Furthermore, the maximum and minimum battery cell voltages are controlled
simultaneously. If the battery cell voltages drop below desired voltage level, the
microcontroller can be programmed to alarm the user via the GUI or a speaker mounted
on the MAX11068.
The Arduino Due has two I2C interfaces which are SDA1 & SCL1 and SDA (pin
#20) & SCL (pin #21). The lowest MAX11068 device of SMBus ladder (Board A)
communicates directly with Arduino Due command module. The other MAX11068
devices in SMBus ladder are communicated with Arduino Due command module through
the device below them via level-shifted I2C communication bus.
In Table 16, the pin mapping of Arduino Due that used in smart battery PMU
system can be found (Arduino).
Table 16. Pin mapping of Arduino Due that used in system
Pin Number
20
21
52
53
GND
Net Association
SDA
SCL
Digital Pin 52
Digital Pin 53
Ground
29
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
3.6.
30
Comparison of Selected BMS Boards
In the Table 17, the EVKits or ICs that can be used for managing lead-acid batteries are
given. The Table 17 is done according to Maxim Integrated Evaluation System for the
MAX11068 (2013), Maxim Integrated MAX17830 12-Channel, High-Voltage Battery
Sensor with Advanced SMBus Ladder and External Cell Balancing (2013), Maxim
Integrated MAX11068 12-Channel, High-Voltage Battery Sensor, Smart DataAcquisition Interface (2010), Maxim Integrated Automotive Product Guide (2012) and
Linear Technology LTC6804-1/LTC6804-2 Multicell Battery Monitors (2013).
Table 17. Comparison of Different BMS Evaluation Kits
Properties
General
Part No
Availability
Cells
Balancing
Voltage Range
(for whole pack
without
external power)
Topology
Series cells,
min
Series cells,
max (no
isolators)
On chip,
passive
External,
passive
IC
Current
Drain
Typical
Standby (uA)
Typical
Operating (mA)
MAXIM INTEGRATED
MAX11068
EVKit
All versions
are not
recommended
for new
designs.
6V to 1.4kV,
approximately
(nominal)
MAX17830
EVKit
Future Product
LINEAR
TECHNOLOGY
LTC6804
Demo Board
Available
9V to 2.48kV,
approximately
(nominal)
11V to 750V,
approximately
(nominal)
Modular
Design: 12
cells/board
4
Modular
Design: 12
cells/board
4
Modular Design: 12
cells/board
372
372
192
External
resistors are
needed
External
resistors and
MOSFETs are
needed
75
External
resistors are
needed
External
resistors and
MOSFETs are
needed
N/A
External resistors
are needed
2
N/A
4
External resistors
and MOSFETs are
needed
35
0.45
3. SYSTEM CONSIDERATIONS BASED ON MAX11068
Readings
Typical Balance
(mA)
Voltage
measurement
accuracy
Cell Volt (V),
min
Cell Volt (V),
max
Temperature
measurement
Data rate
Cost
EVKit
IC only
31
250
N/A
N/A
±0.25%, ≤5
mV Offset
Voltage
0
±0.1%
±0.03%
0
0
5
5
5
Two auxiliary
inputs for 12
cells
10 kHz to 200
kHz
N/A
N/A
Two auxiliary
inputs for 12
cells
N/A
Two auxiliary
inputs for every 12
cells
Up to 1 MHz
N/A
N/A
$150
$15.25
First of all, from the Table 17, we obtain that LTC6804 is the best choice, if the designers
have not had MAX11068 yet, at the moment to design new Battery Management Systems
because it is the only available EVKit or IC.
MAX17830 is Maxim Integrated’s second generation (next-generation version of
the MAX11068) high-accuracy & high-voltage battery-management solution Maxim
Integrated MAX17830 12-Channel, High-Voltage Battery Sensor with Advanced SMBus
Ladder and External Cell Balancing (2013). Since it is mentioned MAX11068 will be no
longer produced (Maxim Integrated Evaluation System for the MAX11068 2013),
MAX17830 could be chosen instead of it for future designs. In our MSc. Thesis project
MAX11068 was used as it was produced before and consequently available, and
MAX17830 is not available at the moment.
Furthermore, MAX11068 has some advantages as well as some disadvantages
over the LTC6804. The voltage range for whole pack without any external power of
MAX11068 is greater than LTC6804. Moreover, max series cells of MAX11068 are
greater than LTC6804. On the other hand, LTC6804 has more voltage measurement
accuracy than MAX11068 and the data rate range of LTC6804 is greater than
MAX11068. Therefore, the designers should select the right BMS EVKit based on their
requirements.
Apart from these, Texas Instrument’s bq34z110 which is a lead-acid fuel gauge
can be the other option to monitor and state the SoC and SoH of a battery (Texas
Instruments 2012). However, it does not have any balancing and protection solution.
Therefore, an additional balancing and protection board is needed. Thus, the functionality
of bq34z110 is less than MAX11068.
32
4.
POWER MANAGEMENT SYSTEM DESIGN
The complete measurement system, which is illustrated in Figure 16, is consisted of four
major parts:
 One 48V battery pack of series-connected 24 lead-acid cells.
 Two MAX11068 evaluation kits (EVKits)
 Arduino Due
 Constant Current Load
Figure 16. The complete measurement system
The connections of lead-acid battery and Arduino Due to the MAX11068 EVKit are
given in Figure 17.
Figure 17. Connection Setup of MAX11068
4. POWER MANAGEMENT SYSTEM DESIGN
33
-
D6 + - D5 + - D4 + - D3 + - D2 + - D1 +
+
C6
-
B6 + - B5 + - B4 + - B3 + - B2 + - B1
+
The series connected 24-cells lead-acid battery design is given in Figure 18.
A6
+ -
A2
C1 -
A1
-
+ -
A3
C2
+
+ + -
A4
C3
+ -
+ + -
A5
C4
+ -
C5
+
-
+ + -
Figure 18. Series connected Lead-Acid battery design
In addition to the wires given in Figure 18, there are also wires which used to
connect the battery cells to MAX11068 EVKits.
From the Figure 19, the connection of J3 Header of MAX11068 Evaluation Kit+
PACK A and Arduino Due can be seen. The serial clock line (SCL), serial data line
(SDA), faster wakeup (FWUP), backup voltage (VDDBU) and ground (GND) pins of
Arduino Due is connected to J3 Header of MAX11068 Evaluation Kit+ PACK A which
provides the I2C communication. Since the SHDNINA net should be pulled up to 3.3V
(Maxim Integrated MAX11068 Evaluation System 2010), digital pins 52 & 53 of
Arduino Due are connected to J3 Header of MAX11068 Evaluation Kit+ PACK A
because the input/output (I/O) pins run at 3.3V.
Figure 19. Arduino Due & J3 Header of MAX11068 Evaluation Kit+ PACK A
Connection
4. POWER MANAGEMENT SYSTEM DESIGN
34
Furthermore, the constant current load is illustrated in Figure 20. When the voltage is
applied on the battery voltage (Vbat), the gate resistor (RG) turns the metal-oxidesemiconductor field-effect transistor (MOSFET) (IRL2910) ON. In addition, the baseemitter voltage (VBE(on)) of NPN transistor BC549 is equal to 0.7V. The relationship
between the sensing resistor (RS) and drain current (ID) is given in the equation (12).
VBE ( on ) 0.7V
(12)
RS 

ID
ID
Hence, ID is controlled by the RS value. In the load, RS is chosen as 0.75Ω. Hence,
the constant current through the load is set as approximately 0.9A. When the potential
difference between RS reaches VBE(on), the transistor is turned ON.
The equation for power dissipation of MOSFET (PD) is given in the equation (13).
It is calculated by multiplication of the drain current with potential difference of drainsource.
(13)
PD  (Vbat  VBE (on))  ID
Nonetheless, a big value such as 10kΩ is chosen for RG so that the transistor is not
damaged.
Vbat
3
RG
10k
IRL2910
BC549B
Q2
2
M1
1
RS
0.75
0
Figure 20. Load design
This constant current load design has some advantages. First, the current can be kept
stable independent of battery voltage. Secondly, the RS dissipates low power since the
potential difference across it is nearly 0.7V.
Consequently, with the help of this constant current load, the load discharging
curve can be observed.
4. POWER MANAGEMENT SYSTEM DESIGN
Lastly, the complete system setup is illustrated in Figure 21.
Figure 21. Complete system setup
The numbers given in Figure 21 denotes that:
1.
2.
3.
4.
5.
6.
7.
8.
Cooling fan for the MOSFET
Constant current load without sensing resistor
MS8221D multimeter (Mastech 2013)
24 cells lead-acid battery pack
Sensing resistor
Connection socket
Arduino Due
Two MAX11068 EVKits
In the complete system setup, the multimeter is used for measuring the current
passing through the load. Also, the connection sockets are connecting two MAX11068
EVKits to the 24 cells lead-acid battery pack. Also, there are two connection sockets for
each 12 lead-acid battery cell for the series connected complete 24-cells lead-acid battery
setup.
35
36
5.
PROGRAMMING OF POWER MANAGEMENT
UNIT
NetBeans IDE is used for Smart Battery Power Management Unit project. NetBeans IDE
is written in one of most widely used programming language, Java (Net Beans 2013).
A Java-based software program, NetBeans IDE, is used in this project for the
voltage measurement and display of each battery cell voltage including maximum and
minimum voltages of cells of the battery pack and helps to establish a connection with
user.
Programming of MAX11068 evaluation kit (EVKit) is the major part of Smart
Battery Power Management Unit (PMU) project. The Programming of Smart Battery
PMU project is mainly divided into three parts which are: Serial Communication,
graphical user interface (GUI) and the main classes.
5.1.
Flowchart of Programming of PMU
At the beginning, the GUI is initialized. After that, universal serial bus (USB) to Serial
Communication is initialized via getting port identifiers and searching a serial
communication (COM) port element. If port enumeration contains more elements it
returns the next element of this enumeration. If COM port is not found; the software
returns error result. On the other hand, if COM port is found; then serial port and USB to
Serial input/output (I/O) streams are opened, the event listeners are added, and the
function checks while the input data is stated as sent and available. When the USB to
Serial Communication is completed, the voltages are acquired. The acquired voltages are
read. After reading voltages, the State of Charge (SoC) calculation function is called.
Thereafter, the measurements are updated in GUI. When the measurements are updated
then the program returns to get more voltage measurements.
Simultaneously, when the measurements are updated in GUI, the software gives
users two choices which are shutting down the software or writing data to Microsoft
Excel Sheet. The data is written to Microsoft Excel Sheet via creating a new excel file
and sheet, instantiating cell objects and adding for the Excel sheet, writing cell voltages,
SoC results, measurement counter and timestamp to file from measurement registers.
When the data is written to Microsoft Excel sheet then the program returns to get more
5. PROGRAMMING OF POWER MANAGEMENT UNIT
voltage measurements. The full simplified and visualized flowchart is illustrated in
Figure 22. The steps that the software follows are shown in the flowchart.
Start
GUI initializations
Initializing USB to
Serial
Communication
Getting voltages
Read voltages
Calculate SoC
Updating the
measurement in GUI
User shuts down the
software
End
Figure 22. The flowchart of programming of PMU
GUI Write to Excel
button
Writing Data to
Microsoft Excel
Sheet
37
5. PROGRAMMING OF POWER MANAGEMENT UNIT
5.2.
38
Java Libraries
The core Java libraries are used for developing Smart Battery PMU which are listed in
Table 18.
Table 18. The used core Java libraries
java.awt
java.util
The core Java libraries
java.io
java.nio
javafx
javax.swing
java.text
jxl
The “java.awt” is utilized for generating user interfaces and graphics. Moreover,
“java.io” is made use of providing system input and output via data streams and
serialization. “java.nio” is used for defining buffers that are containers of data. “java.text”
is used for processing texts and numbers. “java.util” is utilized for handling events, date
and time. In addition, “javafx” is made use of ensuring the modern-looking user interface.
“javax.swing” is used on top of the “java.awt”. Furthermore, “jxl” is utilized for writing,
reading and updating Microsoft Excel sheets dynamically.
5.3.
State of Charge (SoC) Calculation Function
The each battery cell voltage and SoC is set to 0 in code at the beginning of
measurement.
The measurements are being updated by first, and individual Battery Management
System (BMS) cell voltages are acquired and added together. Following this addition, the
total voltages are divided to 24. As a consequence, total battery voltage is calculated.
After that, SoC calculation method is called and those BMS cell voltages are placed to
that method. Consequently, SoC of each cell is calculated. Then, all individual cell SoCs
are added and divided to 24 to find total battery SoC.
39
6.
RESULTS AND DISCUSSION
For testing the Power Management Unit (PMU), the software titled RFIC PMU GUI is
used. A new graphical user interface (GUI) is designed for because it allows to record to
excel, displays State of Charges (SoCs) and it is easy to use.
6.1.
Graphical User Interface
The main GUI window given in Figure 23 displays the voltage and SoC status of 24 cells
of the lead-acid battery pack. The total battery voltage is calculated by adding all series
individual cell voltages. “Write to Excel” button is included to write measurement data to
excel sheet. The PMU GUI is updated instantaneously. The data update rate including the
delay is set by microcontroller.
For each cell there is a status indicator. If that status indicator shows “OK” it
means the cell is in a safe condition. If that status indicator shows “OV” then the cell is
overcharged (the cell voltage is higher than 2.15V). On the other hand, if that status
indicator shows “UV” then the cell is undercharged (the cell voltage is less than 1.7V).
Both the over & under voltage values are set in Java code and cannot be changed in GUI.
Differently from displaying each cell, the GUI also shows the total lead-acid
battery pack voltage and SoC. It also plots total lead-acid battery SoC and voltage versus
time.
Figure 23. PMU GUI without charts
6. RESULTS AND DISCUSSION
40
With this PMU GUI, the users can be aware of what is going on the lead-acid battery
pack. The plots make it easier to obtain the variations by the time of progress. If the
battery or cell status indicator is changed then the hardware should be fixed before any
serious damage is caused. This PMU GUI allows the lead-acid battery for long-life
operation by monitoring the voltages and SoCs of 24 lead-acid battery cells in a smart
way. Only monitoring the voltages and SoCs is not enough for long-life operation.
However, if the voltages and SoCs are not monitored then they could not be controlled.
The program immediately starts when it is run by the user. The only userdependent function of program is “Write to Excel” function.
6.2.
Cell Measurement Results
The Figure 24 shows the individual cell voltages and SoCs together with status indicator
when the load is not connected at a given time.
Figure 24. Individual cell voltages and SoCs
From the Figure 24, we can detect that Cell #19’s voltage status is less than other
cells. The status indicator shows “OK”. However, it is going to enter a critical level
(1.9V) in a time of period if it is not charged. In addition, Cell #12 and Cell #21’s SoC
status’ are less than 40% and they should also be handled well. According to the Figure
24, since not all cells are in a same voltage level, the passive or active balancing might be
applied to obtain the optimal performance of lead-acid battery pack.
Furthermore, all lead-acid cells are in a safe voltage range because none of them
is overcharged or undercharged.
6. RESULTS AND DISCUSSION
41
The SoC icons are defined in GUI as given in Table 19. The intervals are defined based
on Table 5.
Table 19. SoC (%) and SoC Icon with respect to it
SoC (%)
[90-100]
[80-90)
[70-80)
[60-70)
[50-60)
[40-50)
[30-40)
[20-30)
[10-20)
[0-10)
SoC Icon
6. RESULTS AND DISCUSSION
6.2.1.
42
GUI Chart Results
The measurement result at a certain time is removed from the charts after one hour. In
other words, measurement results last for one hour in GUI charts and new measurement
results are shown. The measurement results shown in Chapter 6.2.1 are taken without
using load. In Figure 25, variation of total 24-cells battery SoC in 15 seconds is given.
Figure 25. Total Battery SoC vs. Time
The small fluctuations can be seen in 15 seconds from Figure 25.
Figure 26. Total Battery SoC vs. Time in a different period of time
The whole 24-cells battery SoC remains almost in a same level, 48%, in one hour.
However, in some periods the variations can be observed from Figure 26.
6. RESULTS AND DISCUSSION
43
In the Figure 27, the variation of total 24-cells battery voltage in 15 seconds is given.
Figure 27. Total Battery Voltage vs. Time
Figure 28. Total Battery Voltage vs. Time in a different period of time
From Figure 28, the total 24-cells battery voltage is approximately equal to 48V.
As it is discussed in section 2.4, a SoC calculation algorithm is applied to estimate SoC of
battery cell voltages. In a few words, it is estimation. However, the total battery voltage is
directly shown in GUI charts without any estimation or calculation. Therefore, it is also
obtained from Figures 27 and 28 that the total battery voltage is more stable than total
battery SoC.
6. RESULTS AND DISCUSSION
44
In the Figure 29, the instantaneous total battery voltage, SoC and status are given.
Figure 29. Instantaneous Battery Voltage, SoC and Status
According to the Figure 29, it can be acquired the battery is in a safe condition because
the total battery SoC is more than 40%. The total battery is charged 48% and its voltage
is 48.74V. The clock is also added to GUI so the users can observe battery voltage, SoC
and status at any time.
6.2.2.
Excel Results
The Excel sheet displays the measurement counter, cell voltages, cell SoCs, date and
time. The measurement counter is updated every second.
Figure 30. Cell voltages and measurement counter displayed in Excel sheet
From the Figure 30, the changes of each cell voltages with respect to the
measurement counter can be obtained. The first label is measurement counter. After that,
the results of voltages of 24 cells are located in labels.
6. RESULTS AND DISCUSSION
45
The labels that show SoC results of 24 cells follow the labels that show the voltage
results of 24 cells.
Figure 31. Cell SoC and date & time displayed in Excel sheet
From the Figure 31, the changes of each cell SoC with respect to the date & time
can be obtained.
6.3.
Load Discharging Curve
The load discharging curves help to observe the health of battery cells. The battery cell
discharge performance depends on the load and how they are used. In Figure 32, the
behavior of total battery voltage with respect to time when the load is connected is
illustrated.
The current through the load is 0.74A at the beginning of discharging process. It is
dropped to 0.72A at the end of the discharging process. Therefore, there is a 0.02A
difference between the beginning and end of discharging process.
6. RESULTS AND DISCUSSION
46
Total Lead-Acid Battery Voltage vs Time
50
Total Lead-Acid Battery Voltage (V)
45
40
35
30
25
20
15
10
5
0
0
20
40
60
Time (mins)
80
100
120
Figure 32. Total lead-acid battery voltage versus time
In the Figure 32, the total battery voltage is 50.87V at the beginning. After 116
minutes it is dropped to 44.8V.
Apart from Figure 32, in Figure 33 when the load is connected the load
discharging curve of lead-acid battery cell #19 with respect to time is given.
Load Discharging Curve
2
Lead-Acid Battery Cell #19 Voltage (V)
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
20
40
60
Time (mins)
80
100
120
Figure 33. Load Discharging Curve of lead-acid battery cell #19
6. RESULTS AND DISCUSSION
47
From Figure 33, it is obtained the power delivered by cells with the inclined discharge
curve drops during the discharge cycle. The lead-acid battery cell #19 voltage is 2.09V at
the beginning. After 116 minutes it is fallen to 1.68V. The measurement is stopped when
the least charged lead-acid battery cell (cell #19) is less than 1.7V. Here, 1.7V is the
critical level that states the lead-acid battery cell is fully discharged (Enersys Cyclon
2008).
In addition, since the constant current load is used for the discharging process, the
0.02A variation between the beginning and end is acceptable.
Furthermore, the lead-acid battery cells must be fully charged after every
discharging process and they need to be kept charged at a floating voltage which is more
than the nominal voltage.
6. RESULTS AND DISCUSSION
48
After the discharging process and load is disconnected from the setup, the battery cell
voltages are also measured by the MS8221D multimeter (Mastech 2013). The
comparison between the GUI and multimeter measurements are given in Table 20.
Table 20. Comparison of GUI and multimeter lead-acid battery cell measurements
Cell #1
Cell #2
Cell #3
Cell #4
Cell #5
Cell #6
Cell #7
Cell #8
Cell #9
Cell #10
Cell #11
Cell #12
Cell #13
Cell #14
Cell #15
Cell #16
Cell #17
Cell #18
Cell #19
Cell #20
Cell #21
Cell #22
Cell #23
Cell #24
GUI
1.95
1.96
1.96
1.95
1.95
1.94
1.97
1.97
1.96
1.97
1.97
1.92
1.96
1.97
1.94
1.95
1.96
1.96
1.9
1.95
1.91
1.96
1.95
1.96
Multimeter
1.94
1.95
1.95
1.94
1.94
1.93
1.96
1.96
1.96
1.97
1.96
1.91
1.96
1.96
1.93
1.95
1.95
1.95
1.89
1.95
1.91
1.95
1.94
1.96
From Table 20, it can be concluded that the GUI shows correct results. The 0.5%
difference between the GUI and multimeter measurements is acceptable. The difference
might be caused by rounding, tolerances and/or cable losses.
50
7.
CONCLUSIONS
In conclusion, this MSc. Thesis studies the smart battery PMU for a 48V battery pack of
series-connected 24 lead-acid cells. An alternative voltage and SoC monitoring system is
proposed for current Battery Management Systems which monitors the voltages and
SoCs of lead-acid battery cells properly during both charging and discharging. In
addition, the software allows users to save measurement data to utilize them whenever it
is required. The main goal which is set before the MSc. Thesis is achieved.
During the MSc. Thesis project; at first the initial calculations are performed, after
that initial calculation results are applied in system prototype with minor modifications.
Later on, system is designed with the Java software to control whole PMU system.
The proposed system for the smart battery PMU monitors the individual lead-acid
cell’s performance using the microcontroller board Arduino Due. The Arduino Due
communicates with MAX11068 EVKit through an I2C bus. The Arduino Due provides
communication for measurements of the voltages of each cell and based on those cell
voltage measurements the SoCs are calculated by the Java program. The Arduino Due
commands are written based on Maxim Integrated (2010).
Although this smart battery PMU is designed on a battery back of 24 lead-acid
cells, the complete system can be used for a lead-acid battery pack of any size (up to 372
cells) with minor modifications in the system. The software allows users to make minor
modifications for possible future requirements.
It is proved that the software is measuring lead-acid cell voltages accurately (with
an error margin of 0.5%). The 0.5% difference between the GUI and multimeter
measurements is admissible.
To improve this Smart Battery PMU project, the possible future design is
discussed.
As a consequence, this MSc. Thesis is a part of major KERS project (Tampere
University of Technology RF Integrated Circuits Laboratory 2013) that is going to
interact with the DC/DC converter and motor controller for mobile or stationary high
volume applications. The KERS project requires a 48V battery which provides energy to
DC/DC converter and motor controller. Also, the DC/DC converter and motor controller
restore energy to 48V battery. Hence, this MSc. Thesis can be used for the bigger KERS
project as it is a 48V smart battery PMU which allows providing and restoring energy,
and provides an accurate voltage and SoC monitoring system.
50
8.
CONTINUATION DESIGN OF WORK
The whole system is done by RFIC Laboratory Research Group. The future BMS/PMU
block diagram plan is designed and illustrated in Figure 34. It will be mainly consisted of
source, control and load parts. In this MSc. Work, I work in the monitoring part of the
system. In Figure 34, the communications between 48V 24-cell lead-acid battery,
microcontroller (Arduino) and MAX11068 EVKits in control part are already done. My
MSc. work can be improved by the block diagram given in Figure 34.
BMS BLOCK DIAGRAM JUHA & SELIM / Rev.1.1 / 26-04-2013
SOURCE
CONTROL
LOAD
220V AC
1.2A
LAPTOP
CHARGER
12/24/36V
3A
48V
3A
DCDC BOOST
LTC3862
(DC1286A-A)
NMOS
BUK654R075C
48V
3A
48V
NMOS
BUK654R075C
48V 24-Cell
Lead-Acid
Battery
GND
LOAD
LEDs
13
48V
0.1A
BMS
MAX11068
EVKIT
I2C+Control
DCDC BUCK
LTC3890
(DC1804A)
10
12V
GATE1
GATE
DRIVER 1
IR2117
LDO
LM7810
(OPTIONAL)
GATE
DRIVER 2
IR2117
GATE2
10V
5V
µC
Arduino Uno
5V
4
USB
Figure 34. Future BMS/PMU Block Diagram
The source part is consisted of charger, step-up DC/DC converter and metaloxide-semiconductor (MOS) switch. The charger is selected so that the output of charger
should be 12V, 24V or 36V. LTC3862 is placed at the output of charger, which is a stepup DC/DC converter (Linear Technology 2008). The output of LTC3862 is 48V with a
current of 3A. Two BUK654R0-75Cs are used as MOS switches.
The control part is consisted of gate drivers, 24-cell lead-acid batteries, MAX11068
evaluation kit (EVKit) (as a BMS), Arduino Uno (as a microcontroller), and an optional
51
low-dropout regulator (LDO). In addition, Arduino Due can be replaced with Arduino
Uno. The major difference for this design is that Arduino Due’s operating voltage is
3.3V. Two IR2117s are planned to use as high side gate drivers and for level shifting
purpose. However, there are some limitations to use them. IR2117s have a bootstrap
capacitor that pulls the logic supply voltage into gate node when it turns on. The supply
voltage charge in capacitor lasts for only couple milliseconds. Next, another pulse is
needed for input pin. Hence, IR2117s can be only used with a 1+ kHz pulse width
modulation (PWM) signal. Thus, two IR2117s can be replaced in future. High side gate
driver configuration needs a switch between the load and supply. The MOS gate is pulled
high to turn on the high side gate driver. With the usage of them, the Arduino Uno can
control the switches both on source and load sides. Otherwise, the Arduino Uno cannot
control the switches as its operating voltage is 5V. Furthermore, the Arduino Uno
requires input voltage between 7V to 12V to operate. This operating voltage is gained
through a DC/DC step-down process. For this purpose, LTC3890 DC/DC step-down
converter is added to design. LTC3890 takes the 48V input voltage and gives 12V output
voltage (Linear Technology 2010). Additionally, the MAX11068 EVKit is used as BMS.
The inter-integrated circuit (I2C) communication and control of MAX11068 is provided
by the microcontroller. The LDO is optional but it provides filtering and drops out the
output voltage of LTC3890. LM7810 can be used as a LDO in this design to drop 12V
out to 10V (Fairchild Semiconductor 2013).
The load part is consisted of MOS switch and the light-emitting diodes (LEDs)
which are used as a load. The load is connected to MOS switch and the output of 24-cell
lead-acid battery.
Nevertheless, the future BMS/PMU can display temperature and current
measurements of each battery cells. In addition, the portability of the system could be
increased with researching wireless capabilities. The experiments are done with 24 leadacid battery cells. However, considering the MAX11068 specifications in future the
system can be designed to be scalable up to 372 cells.
Nonetheless, an alarming function could be added to the BMS/PMU. The
complete BMS/PMU can make use of the alarming function if any system parameters
exceed their operating levels. Also, if the number of series connected battery cells create
a high voltage level for applications then the cell balancing should be implemented in
future design to prevent the potential damage. The active balancing method can be
applied for this purpose because the active balancing method is faster and the power
efficiency of it is higher than the passive balancing method.
Consequently, the control of the PMU can be mainly divided into four parts:
Avoidance of over/undercharging of battery cells, monitoring battery temperature,
monitoring battery output current, and controlling the switching process.
52
REFERENCES
Andrea, D. 2010. “Battery Management Systems for Large Lithium-Ion Battery Packs”.
Artech House.
Arduino. “Arduino Due”. [Accessed on 18.04.2013]. Available at:
http://arduino.cc/en/Main/arduinoBoardDue
Arduino. “SAM3X-Arduino Pin Mapping”. [Accessed on 18.04.2013]. Available at:
http://arduino.cc/en/Hacking/PinMappingSAM3X
Arduino. “Arduino Uno”. [Accessed on 18.04.2013]. Available at:
http://arduino.cc/en/Main/arduinoBoardUno
Arduino. “Wire Library”. [Accessed on 18.04.2013]. Available at:
http://arduino.cc/en/reference/wire
Arendarik, S. 2012. “Active Cell Balancing in Battery Packs”. Freescale Semiconductor,
Inc. [Accessed on 18.04.2013]. Available at:
http://www.freescale.com/files/32bit/doc/app_note/AN4428.pdf
Atmel. “SAM3X/A”. [Accessed on 18.04.2013].Available at:
http://www.atmel.com/Images/doc11057.pdf
Begovic M. 2012. “Electrical Transmission Systems and Smart Grids”. Springer.
Bergveld, H.J. Kruijt, W.S. Notten, P.H.L. 2002. “Battery Management Systems –
Design by Modelling”. Kluwer Academic Publishers. Boston.
Broussely, M. Pistoia, G. 2007. “Industrial Applications of Batteries”. First Edition.
Elsevier B.V. Amsterdam, The Netherlands.
53
Clarke, R.S. 2011. “B-LAB: CHANGING THE GAME FOR LEAD ACID”. Applied
Intellectual Capital. Alameda. [Accessed on 24.09.2013]. Available at:
http://www.battcon.com/PapersFinal2011/ClarkeStephenPaperDONE2011_13.pdf#!.
Claus, D. 2008. “Materials and processing for lithium-ion batteries”.Volume 60, Issue 9,
pp 43–48.
Enersys Cyclon. 2008. “US-CYC-ApplicationManual-007_04081”.
Fairchild Semiconductor. 2013. “LM78XX / LM78XXA 3-Terminal 1 A Positive
Voltage Regulator”. [Accessed on 19.04.2013]. Available at:
http://www.fairchildsemi.com/ds/LM/LM7805.pdf
Huet, F.1998. “A review of impedance measurements for determination of the state-ofcharge or state-of-health of secondary batteries”. Elsevier Science S.A.
Hyperion. “HYPERION AC/DC EOS0606iAD-C USER’S MANUAL”. [Accessed on
23.04.2013]. Available at:
http://media.hyperion.hk/dn/EOS/EOS0606iAD(acdc)-A-MAN-EN.pdf
International Rectifier. 2007. “IR2117(S)/IR2118(S) & (PbF)”. [Accessed on
20.04.2013]. Available at: http://www.irf.com/product-info/datasheets/data/ir2117.pdf
Kong-Soon, N. Chin-Sien, M. Yi-Ping, C. Yao-Ching, H. 2008. “State-of-Charge
Estimation for Lead-Acid Batteries Based on Dynamic Open-Circuit Voltage”. 2nd IEEE
International Conference on Power and Energy (PECon 08). Johor Baharu, Malaysia.
Krein, P. T. Balog, R. 2002. “Life Extension Through Charge Equalization of Lead-Acid
Batteries”. [Accessed on 18.04.2013]. Available at:
http://windandsunpower.com/Download/intelec02.pdf
Linden, D. Reddy, T. B. 2002. “HANDBOOK OF BATTERIES”. Third Edition.
McGraw-Hill. New York.
Linear Technology. 2008. “LTC3862 Multi-Phase Current Mode Step-Up DC/DC
Controller”. [Accessed on 20.04.2013]. Available at:
http://cds.linear.com/docs/en/datasheet/3862fb.pdf
54
Linear Technology. 2010. “LTC3890 60V Low IQ, Dual, 2-Phase Synchronous StepDown DC/DC Controller”. [Accessed on 19.04.2013]. Available at:
http://cds.linear.com/docs/en/datasheet/3890fc.pdf
Linear Technology. 2013. “LTC6804-1/LTC6804-2 Multicell Battery Monitors”.
Mastech. 2013. “MS8221D DIGITAL MULTIMETER”. [Accessed on 17.12.2013].
Available at: http://www.p-mastech.com/images/SPEC/ms8221d%20ms8230b.pdf
Maxim Integrated. 2008. “Maxim MINIQUSB User Guide”. [Accessed on 04.04.2013].
Available at: http://datasheets.maximintegrated.com/en/ds/MINIQUSB.pdf
Maxim Integrated. 2010. “MAX11068 12-Channel, High-Voltage Battery Sensor, Smart
Data-Acquisition Interface”.
Maxim Integrated. 2012. “Automotive Product Guide”. [Accessed on 04.10.2013].
Available at:
http://digichip.ru/datasheet/PDF/708f64cd8fd71531242118389f171796/6cc98ba2045fe40
98d24356d14447085/MAX2135AETNV+.pdf. p. 29.
Maxim Integrated. 2013. “Evaluation System for the MAX11068”. [Accessed on
04.04.2013]. Available at:
http://www.maximintegrated.com/datasheet/index.mvp/id/6882.
Maxim Integrated. 2013. “MAX17830 12-Channel, High-Voltage Battery Sensor with
Advanced SMBus Ladder and External Cell Balancing”. [Accessed on 04.10.2013].
Available at: http://www.maximintegrated.com/datasheet/index.mvp/id/7114.
Miller, P. 2012. “xEV market trend and prospect”. 2012 IEEE Vehicle Power and
Propulsion Conference. Seoul, Korea.
National Geographic 2004. “The End of Cheap Oil”. vol. 205, no. 6.
NetBeans. 2013. “Welcome to the NetBeans Community”. [Accessed on 09.05.2013].
Available at: https://netbeans.org/about/index.html
NXP Semiconductors. “UM10204, I2C-bus specification and user manual”. Rev. 5-9
October 2012.
55
Pang, S. Farrell, J. Du, J. Barth, M. 2001. “Battery State-of-Charge Estimation”.
Proceedings of the American Control Conference, vol. 2.
Pistoia, G. 2005. “Batteries for Portable Devices”. Elsevier B.V. Amsterdam, The
Netherlands.
Pistoia, G. 2009. “Battery Operated Devices and Systems”. First Edition. Elsevier B.V.
Amsterdam, The Netherlands.
Racherla, K. 2011. “Measure Multiple Temperatures in Battery-Management Systems,
and Save Power Too”. [Accessed on 07.10.2013]. Available at:
http://www.eetasia.com/STATIC/PDF/201205/EEOL_2012MAY03_POW_TEST_AN_0
1.pdf?SOURCES=DOWNLOAD
Rahimi-Eichi, H., Ojha, U., Baronti, F., Chow, M. 2013. "Battery Management System:
An Overview of Its Application in the Smart Grid and Electric Vehicles," Industrial
Electronics Magazine, IEEE, vol.7, no.2.
Reddy, T. B. 2011. “Linden’s Handbook of Batteries”. Fourth Edition. McGraw-Hill.
New York.
Tampere University of Technology RF Integrated Circuits Laboratory. 2013. “Mobile /
Stationary Big Volume Applications”. [Accessed on 09.01.2014]. Available at:
http://www.cs.tut.fi/tlt/rfasic/RFCClab.pdf
Tanaami, A. Morimoto, M. 2009. “On-line Estimation of SOH for Lead-Acid Battery”.
PEDS 2009. International Conference on Power Electronics and Drive Systems. Taipei,
Taiwan, R.O.C.
Texas Instruments. 2006. “bq20z80 Application Book”.
Texas Instruments. 2012. “Wide Range Fuel Gauge with Impedance Track™ for LeadAcid Batteries” [Accessed on 04.10.2013]. Available at:
http://www.ti.com/lit/ds/slusb55a/slusb55a.pdf
The Boston Consulting Group Inc. 2010. “Batteries for Electric Cars; Challenges,
Opportunities, and the Outlook to 2020”. Boston, MA, USA. [Accessed on 19.04.2013].
Available at: http://www.bcg.com/documents/file36615.pdf
56
Wang, P., Liang, B., Ye, X., Ko, W.H., Cong P. 2010. “A Simple Novel Wireless
Integrated Power Management Unit (PMU) for Rechargeable Battery-Operated
Implantable Biomedical Telemetry Systems,” Bioinformatics and Biomedical
Engineering (iCBBE), 4th International Conference. Chengdu.
Wikipedia. 2009. “Grid storage energy flow”. [Accessed on 12.04.2013]. Available at:
http://en.wikipedia.org/wiki/File:Grid_storage_energy_flow.png.
Williams, D. 2012. “Configuring the bq34100 Data Flash”. Texas Instruments.
57
APPENDIX 1: MAX11068 EVKIT CELL
CONFIGURATION HEADERS AND SWITCH
MAX11068 EVKit’s cell configuration header J101 is given below (Maxim Integrated
2013).
58
MAX11068 EVKit’s cell configuration header J102 is given below (Maxim Integrated
2013).
59
MAX11068 EVKit’s cell configuration switch SW101 is given below (Maxim Integrated
2013).
60
APPENDIX 2: MAX11068 EVKIT CELL
CONNECTION SWITCHES AND HEADERS
MAX11068 EVKit’s cell connection switches and headers are given below (Maxim
Integrated 2013).
61
APPENDIX 3: DISCHARGING ENERSYS CYCLON
BATTERIES
In the table below, the recommended EODV with respect to the Discharge Rate is given
(Enersys Cyclon 2008).
Discharge Rate (A)
0.05C10
0.1C10
0.2C10
0.4C10
C10
2C10
4C10
Minimum EODV/cell (V)
1.75
1.7
1.67
1.65
1.6
1.55
1.5
62
APPENDIX 4: TIMING OF PARAMETER UPDATES
DURING THE RELAXATION MODE FOR STATE OF
HEALTH ESTIMATION
During the relaxation mode, the timing of DoD0 and Qmax updates are given below (Texas
Instruments 2006).